Cigarette smoking among youth has shown considerable fluctuation since the mid-1970s, having first fallen, then remained stable, then raised again, and most recently declined. Moreover, the trends differ strikingly across race and ethnic groups, with larger declines occurring among African-Americans and Hispanics than whites. Given the harm of smoking and the likelihood of teen smokers continuing the habit for decades to come, such trends have generated much interest. However, the research literature has done more to describe the trends than to explain them. Despite speculation about the possible role of changes in individual characteristics that promote smoking and of macro-level changes in cultural norms, social strain, media influence, and cigarette prices, tests of such claims are rare. This project proposes to extend the existing literature on youth smoking by examining both micro and macro-level determinants of changes in youth smoking. It will consider micro-level factors by examining changes in smoking and the individual-level determinants of smoking with survey data on high school seniors from 1976-2001. It will consider macro-level factors by linking the survey data to published data on aggregate determinants of smoking that are measured for the U.S. youth population as a whole. Combining the micro and macro-level data in multilevel models will then allow tests of the explanations at both levels. The goals of the project thus involve secondary analysis of data relevant to tobacco use etiology, and the impact of socioeconomic, ethnic, and cultural factors on cancer-promoting behavior. The aims further reflect a change in the principal investigator's interests to behavioral research in cancer control.